A hallmark of biological systems is the high degree of interactions among their constituent components. Cells can be described as complex circuits consisting of a network of interacting molecules. Key component of these networks are proteins that serve to couple cellular functions. A protein that couples functions can be described as a “molecular switch.” In most general terms, a molecular switch recognizes an effector (input) signal (e.g., ligand concentration, pH, covalent modification) with resultant modification of its output signal (e.g., enzymatic activity, ligand affinity, oligomeric state). Examples of natural molecular switches include allosteric enzymes that couple concentration of effector molecules with level of enzymatic activity, and ligand-dependent transcription factors that couple ligand concentration to output level of gene expression. Molecular switches can be “ON/OFF” in nature or can exhibit a graded response to a signal.
There is recognition that there is great potential to design fusion proteins that act as molecular switches to modulate or report on biological functions for a variety of applications including biosensors (Siegel and Isacoff 1997; Baird, Zacharias et al. 1999; Doi and Yanagawa 1999; de Lorimier, Smith et al. 2002; Fehr, Frommer et al. 2002) modulators of gene transcription and cell signaling pathways (Rivera 1998; Guo, Zhou et al. 2000; Picard 2000), and novel biomaterials (Stayton, Shimoboji et al. 1995). Despite its great potential, however, molecular switch technology has not been extensively exploited, in part due to technical challenges in engineering effective molecular switches. In general, existing approaches to creating protein molecular switches include: control of oligomerization or proximity using chemical inducers of dimerization (CID); chemical rescue; fusion of the target protein to a steroid binding domain (SBD); coupling of proteins to nonbiological materials such as ‘smart’ polymers (Stayton, Shimoboji et al. 1995; Ding, Fong et al. 2001; Kyriakides, Cheung et al. 2002) or metal nanocrystals (Hamad-Schifferli, Schwartz et al. 2002); and domain insertion.
The approach of control using a chemical inducer of dimerization (CID) utilizes a synthetic ligand as the CID that controls the oligomeric or proximity of two proteins (Rivera 1998). CIDs are small molecules that have two binding surfaces that facilitate the dimerization of domains fused to target proteins. This approach was first developed using the immunosupressant FK506 to facilitate dimerization of target proteins fused to the FK506-binding protein, FKBP12 (Spencer, Wandless et al. 1993). Several variations on this system have since appeared as well as a system using the antibiotic coumermycin to dimerize proteins fused to B subunit of bacterial DNA gyrase (GyrB) (Farrar, Olson et al. 2000). CIDs have been used to initiate signaling pathways by dimerizing receptors on the cell surface, to translocate cytosolic proteins to the plasma membrane, to import and export proteins from the nucleus, to induce apoptosis and to regulate gene transcription (Bishop, Buzko et al. 2000; Farrar, Olson et al. 2000). However, CIDs have only been applied to those functions that require changes in the oligomeric state or proximity of the two proteins. As described in the literature however, this approach cannot be readily applied to a single protein.
Chemical rescue has recently been applied as a strategy for control, in the case of dimerization (Guo, Zhou et al. 2000). Chemical rescue aims to restore activity to a mutant, catalytically defective enzyme by the introduction of a small molecule that has the requisite properties of the mutated residues. Since first described for subtilisin (Carter and Wells 1987), chemical rescue has been demonstrated for a number of different mutated protein-small molecule pairs (Williams, Wang et al. 2000). The vast majority of these rescues required >5 mM concentrations to show detectable rescue, and the maximum fold improvement in activity of the mutant was generally less than 100-fold and required >100 mM concentrations of the rescuing molecule.
For the strategy of fusion to a steroid binding domain, the protein to be controlled is fused end-to-end to a SBD (Picard 2000). In the absence of the steroid that binds to the SBD, it is believed that a Hsp90-SBD complex sterically interferes with the activity of the protein fused to the SBD. The disassembly of the complex upon steroid binding restores activity to the protein. This strategy has been successfully applied principally to transcription factors and kinases (Picard 2000). Artificial transcription factors (such as GeneSwitch™) have been developed using this strategy and have promise for tissue-specific gene expression in transgenic animals and human gene therapy (Burcin, B W et al. 1998; Burcin, Schiedner et al. 1999).
For approaches involving coupling to non-biological materials, the protein to be controlled is coupled to a non-biological material that responds to an external signal and thereby affects the protein coupled to it. ‘Smart’ polymers that change their conformation upon a change in pH or temperature have been conjugated to proteins near ligand binding sites, to create switches that sterically block access to the binding site at, for example, higher temperatures, but not at lower temperatures (Stayton, Shimoboji et al. 1995; Ding, Fong et al. 2001). Inductive coupling of a magnetic field to metal nanocrystals attached to biomolecules resulting in an increase in local temperature thereby inducing denaturation, has so far only been applied to DNA (Hamad-Schifferli, Schwartz et al. 2002).
Relatively few studies have attempted to create a molecular switch using the approach of insertional fusion, in which one gene is inserted into another gene. Insertions result in a continuous domain being split into a discontinuous domain. The first example of successful insertion of one protein into another was of alkaline phosphatase (AP) into the E. Coli outer membrane protein MalF, constructed as a tool for studying membrane topology (Ehrmann, Boyd et al. 1990). High levels of alkaline phosphatase activity were obtained in the fusions despite the fact that alkaline phosphatase requires dimerization for activity. Other examples of proteins that have been inserted into other proteins include green fluorescent protein GFP) (Siegel and Isacoff 1997; Biondi, Baehler et al. 1998; Kratz, Bottcher et al. 1999; Siegel and Isacoff 2000), TEM1 β-lactamase (Betton, Jacob et al. 1997; Doi and Yanagawa 1999; Collinet, Herve et al. 2000), thioredoxin (Lu, Murray et al. 1995), dihydrofolate reductase (Collinet, Herve et al. 2000), FKBP12 (Tucker and Fields 2001), estrogen receptor-α (Tucker and Fields 2001) and β-xylanase (Ay, Gotz et al. 1998).
In studies of insertions into GFP, molecular sensors were created by inserting β-lactamase into GFP by random mutagenesis, to create a protein whose fluorescence increased 60% upon binding of the β-lactamase inhibitory protein. Insertions of calmodulin (a Ca2+ binding protein) into GFP resulted in a fusion whose fluorescence changed up to 40% upon increases in Ca2+ concentration (Baird, Zacharias et al. 1999). In a related strategy, the gene for a circularly permuted GFP was sandwiched between the gene for calmodulin and its target peptide M13 to create a series of sensors whose fluorescence intensity increased, decreased or showed an excitation wavelength change upon binding Ca2+ (Nagai, Sawano et al. 2001).
With the exception of the domain insertion strategy, all of the above-described approaches to engineering a molecular switch are limited in the sorts of signals that can be employed or the types of proteins that can be controlled. CIDs have only been applied to those functions that require changes in the oligomeric state or proximity of the two proteins and thus cannot be used to control a single protein. The chemical rescue approach is limited by the inability to apply the method to any desired signal and by the lack of sensitivity (high concentrations of the signal are required for a small change in activity). The SBD strategy appears to be limited as a general method for controlling any protein due to the apparent requirement for end-to-end fusion.
The domain insertion strategy is a promising and generally applicable approach to engineering a molecular switch. However existing domain insertion strategies are limited by the number of possible insertional fusions between the two domains. Generally, methods for generating molecular switches have not provided a systematic way to generate very large numbers of fusions of different geometries that would be ideal for generating and optimizing functional coupling of protein domains in molecular switches.